Confused about Logging hyper parameters

[minwang-ai]

Does anyone know how to log hyper parameters? The documentation is a little bit unclear.                                  https://pytorch-lightning.readthedocs.io/en/stable/extensions/logging.html
When training a model, it’s useful to know what hyperparams went into that model. When Lightning creates a checkpoint, it stores a key “hyper_parameters” with the hyperparams.
lightning_checkpoint = torch.load(filepath, map_location=lambda storage, loc: storage)
hyperparams = lightning_checkpoint["hyper_parameters"]

[scottire]

I'm not sure if this is what you're looking for but PL has a bunch of external loggers which save hyper-parameters. I work for W&B, and our logger can be used to track hyper-parameters, log metrics / models and even log media such as images or text.

https://pytorch-lightning.readthedocs.io/en/stable/extensions/generated/pytorch_lightning.loggers.WandbLogger.html#pytorch_lightning.loggers.WandbLogger

[Blaizzy]

Hey @MinWang1997,

Prince Canuma here, a Data Scientist at Neptune.ai,

There are multiple ways and tools to help you manually or automatically log your experiments(metrics, hyperparameters, model checkpoints and etc) that work with PTL.

Are you currently using any experiment tracker/logger at the moment in your workflow?

I also noticed that the logging hyperparameters section of the PTL docs page you linked to is not very clear, but if you tell me which experient tracker you are using it will be easier for me to help you find the answer.

Here is a list of experiment trackers/loggers that are supported by PTL: https://pytorch-lightning.readthedocs.io/en/stable/extensions/logging.html#supported-loggers

Check it out and let me know which one you are using 😃 !

[Blaizzy]

Hey @MinWang1997,

Any updates?